Radiated emissions testing is only one part of emissions testing. CISPR 32, the document that lays out limits and methods of measurement for emissions (radiated and conducted) for Information Technology Equipment (ITE), broadcast receivers and multimedia equipment is 120 pages in length. ANSI C63.4 deals with only the testing part (limits are contained in a separate regulation) and likewise is a long (and getting longer in the next version) document. Other standards may provide different measurement techniques. This article will deal only with radiated emissions, and then only in a summary fashion, as it relates to CISPR 32 (and ANSI C63.4).
Radiated emissions testing can be broken down into three basic parts: first, is a preliminary test, second is a final test, and the last part is data reduction and comparison with the applicable limits.
This test (or series of tests) is designed to serve a few different purposes. The first is to determine the appropriate Equipment Under Test (EUT) configuration and layout in the test laboratory. The second is to determine the final configuration of the system under test and possibly frequencies to be measured in the next step (final test). The third is part of the first – determine the size of the system and location of the antenna so that the measurement distance can be determined per the standard.
The EUT configuration is one that allows the EUT to function as intended. This may be the EUT by itself in the case of a simple EUT, or it may require the use of additional equipment and cables to allow the EUT to exercise various ports relating to the EUT. These may include ports, such as those needed for external devices such as a monitor, printers, or other devices. The EUT may be a module, in which case there are specific requirements on how it is to be incorporated into a complete system in order to be tested.
CISPR 32 uses a system called “ports” to determine what tests need to be performed. These are different than what a user may regard as ports, in that the EUT as a whole may be regarded as a port (enclosure port), where the user does not think of the enclosure as being a port. This is something that the laboratory or the organization performing the test must be aware of.
For radiated emissions testing (the only port called out for this test is the “enclosure port”) this test would be performed using a less rigorous set of antenna heights and turntable positions. This would be repeated as necessary to find the worst-case configuration for the EUT (arrangement, cable positions, etc.) and to potentially identify the frequencies to be measured in the final test. Where some regulators only require the six worst-case signals to be reported, others may want more. The reader should be aware of the requirements in every country in which the product will be marketed.
CISPR 32 does not call out a separate preliminary test and a final test. However, it may be necessary to consider both, especially if the lab is such that a 3-meter RF semi-anechoic chamber is used for preliminary testing from 30 MHz to 1 GHz and a 10-meter Open Area Test Site (OATS) is used for final testing over this frequency range. Preliminary testing, and a version of the final tests, were performed in a 3-meter RF semi-anechoic chamber and final tests were performed at a 10-meter OATS. If a signal was blocked by an ambient signal at the OATS, then the 3-meter reading in the chamber was used for that frequency. Typically the 20 or so worst frequencies found in the chamber were re-measured at the OATS. As one backed off from 3 meters (in the chamber) to 10 meters (OATS) the relative positions of the signals might change, so the 6 worst in the chamber might not be the 6 worst at the OATS. Twenty frequencies were found to be more than enough to ensure that the 6 worst were captured at the OATS.
How do you determine the measurement distance? Figures C.1 and C.2 of CISPR 32 show sample systems under test where 3 units are labeled as “AE/EUT”. AE means “Auxiliary Equipment”, not the EUT, but required to allow the EUT to function. The figure labels all three (there may be more or less) as AE/EUT so that no particular device must be in one specific location. A circle is drawn that encloses all the AE/EUT boxes and interconnecting cables. This circle is considered to be the boundary of the EUT and is the point from which the distance to the antenna is measured. Note that the center of this circle is the center of the turntable, not the center of the system containing the EUT. The figure in CISPR 32 shows it as the center of one of the boxes on the turntable. This may or may not be the actual case. Where on the antenna do you measure from to determine the measurement distance? From the reference point used for the antenna calibration. This point is typically clearly marked on a log periodic dipole array antenna and generally is the center point of a biconical antenna.
Note that the measurement does not necessarily have to be performed at 3, 5, or 10 meters. There is a formula that allows computation of the limit at distances other than the distances shown in Tables A.2 to A.7 (assuming that the test facility has been validated at the distance to be used). This new limit is given using:
L2 = L1 + 20 log(d1/d2)
L2 is the new limit at distance d2
L1 is the given limit at distance d1
When using this formula, the test report shall show the limit L2 and the actual measurement distance d2. Measurements shall be performed at the 10-meter distance (up to 1 GHz) and the 3-meter distance (above 1 GHz) whenever possible and shall be used as the basis for calculations of limits at other measurement distances.
Another issue to deal with during preliminary testing is the EUT cycle time and the measurement dwell time. The dwell time normally would be longer than the EUT cycle time but may be limited to 15 seconds.
Annex E (which is informative, not normative) provides a list of issues that should be considered during a prescan test (called a preliminary test in this article).
This test may be performed in a qualified RF semi-anechoic chamber or a qualified OATS or FSOATS, depending on the frequencies to be measured. Regardless of the frequency range, the following is appropriate.
The system under test, including the EUT, auxiliary equipment (AE) and cables shall be configured as shown during the preliminary tests to be the worst case. Any software needed to exercise the EUT during the test needs to be loaded and run.
Two options exist for the test lab during the final test. The EUT may be tested over the entire frequency range for the test and the 20 or more worst-case signals fully measured. The second option will likely be used when a 10-meter RF semi-anechoic chamber isn’t available, but a 3-meter chamber is available, along with a 10-meter OATS. In this case, a full final test would be performed in the chamber at 3 meters and the final 20 or so frequencies found during this test would be re-measured at the OATS. Only re-measuring the final 6 frequencies at the OATS would be a poor choice as the worst 6 signals found at 3 meters might not be the same as the 6 found at the OATS. The author recommends that if a full test is performed at an OATS that 10 or more signals be reported, rather than the 6 that some regulators call out as this provides a safety margin for the lab. The same goes for testing in a 3-meter RF semi-anechoic chamber, followed by testing at the worst cast frequencies at 10 meters at an OATS. 20 or more frequencies are recommended in this case for the reason pointed out above and the fact that one or more of the frequencies identified in the chamber may be blocked by ambient signals at the OATS. The author found a number of years ago that a certain regulator “required” measurements at 10 meters, but would accept a few measurements at 3 meters when the OATS data was blocked by ambient signals. Having more than 6 signals reported eased this relaxation of the requirements.
A final test, regardless of the test facility requires that the EUT be fully exercised by the software and that the EUT’s emissions be fully evaluated over the full range of turntable positions (0 to 359 degrees), antenna heights (1 to 4 meters) and both polarities (horizontal and vertical).
DATA REDUCTION AND COMPARISON WITH THE LIMITS
The first part of this section (data reduction) needs to be performed for each signal during preliminary and final testing. While the receiver reads the voltage at its input terminal, the limit (discussed later) is a field strength. The conversion needs to consider the loss in the cables, any pre-amplifier gain, and any antenna factors that convert field strength to voltage. These are simply added together as follows:
Field Strength dBiV/m = Received Voltage dBuV) + Cable Loss dB – Preamplifier Gain dB + Antenna Factor dB/m
Where field strength is the final result, received voltage is the voltage at the input of the receiver, cable loss is the frequency-dependent loss over the length of all cables used in the measurement, preamplifier gain is the frequency-dependent gain of the preamplifier (if used) and the antenna factor is the frequency-dependent factor for that antenna that allows mathematical conversion of the received signal strength to the voltage at the antenna terminals.
This final field strength would be recorded along with the distance between the antenna and EUT, turntable position, antenna height, and polarity for each final signal received from the EUT.
The current edition of CISPR 32 requires that the laboratory compute their measurement instrumentation uncertainty and compare it with a value provided in CISPR 16-4-2 as amended. If their computed measurement instrumentation uncertainty is less than or equal to that value provided in CISPR 16-4-2 (called U<sub>CISPR</sub> in the standard) then the limit provided in Tables A.1, A.2, A.3, A.4, A.5, A.6 or A.7 as applicable are compared against the recorded data directly and a pass/fail determination is made. If the lab’s computed measurement equipment uncertainty is greater than the value of U<sub>CISPR</sub> provided in CISPR 16.4.2 as amended, then the difference between the lab’s computed measurement uncertainty and U<sub>CISPR</sub> is added to the recorded data and that result is compared with the limit that is applicable to the test and a pass/fail determination is made. U<sub>CISPR</sub> is a value that any reasonably equipped and competent laboratory should be able to meet, so this shouldn’t be a problem that the laboratory and customer would have to deal with, but it is a point that both should be aware of. ANSI C63.4 (used in the US) does not have a similar requirement.
This summary should give the reader a starting point and understanding of the basic requirements for radiated emissions testing. While it is based on CISPR 32 and ANSI C63.4, the other standards are similar and this should provide the reader with a starting point. Read the applicable standard to your product and follow it. If there are any differences found, follow the applicable standard for your product.